Chris Day interview.......................
For those who prefer, the audio...................
javascript:DoPlayAudio(document.audioplayer,'http://ds.audiobase.com/projects/realcomp0001/4-047_guest3.jsd','realcomp0001-RC06')
realcomputing.com
--------------------------------------------------------------------------------
Guest: Chris Day
DVORAK: If you look at video, modern video, if you have a satellite dish, or if you've got one of those DVD players, or even if you're looking at video on your computer, many times you find that there's a chip in there from a company called C-Cube Microsystems that is an MPEG decoder chip, that does the job probably better than anybody else's chip. And to discuss that and other issues with modern video, we've got Chris Day, Senior Director of Marketing for the PC-Kodak Division of C-Cube Microsystems from Milpitas, right here in the studio. Chris, welcome to Real Computing.
CHRIS DAY: Thank you very much.
DVORAK: So tell us exactly what the product line looks like.
DAY: C-Cube is the largest supplier of silicon chips for video compression and decompression.
DVORAK: And we're talking about MPEG.
DAY:Yeah, we're talking about MPEG. MPEG being the video standard for a lot of things, like DVD and video broadcasting. We're also talking about DV, and DV is the format that's been adopted by the latest camcorders, digital camcorders.
DVORAK: That's compressed signals... or I guess they're just digital signals that have been somewhat compressed, or how does that work on a DV signal? What's on that tape?
DAY: The DV is... well, C-Cube's advantage here is that we can handle both MPEG and DV. So what we would do from a DV camcorder is be able to take that DV signal, which is a 25 or 50 megabits digital video format, and then be able to edit it. So you'll see C-Cube chips on products from companies like Matrox and Pinnacle and Fast, who supply the video editing market. And what we enable them to do is to go into that video stream and take two digital video streams and combine them together, so you can do video editing in the digital domain.
DVORAK: And that's hard to do.
DAY:That's very hard to do.
DVORAK: Now, what is the current status of MPEG, anyway, in the market. And why don't you explain to people what MPEG, you know, we have MPEG-1, MPEG-2, MPEG-3, MPEG-4.
DAY: Right. Most of the world we're doing is with MPEG-2, we also do MPEG-1. MPEG-1 is a lower resolution standard. Quality-wise it's somewhat similar to the VHS tape that you may have in your house. The biggest market for MPEG-1 is actually in China, where people have these video CD players. So it looks just like a CD, and instead of just providing audio, it's proving video and audio. And that is the largest MPEG-1 market. MPEG-2 is higher quality. If you have a digital broadcast satellite reception in your house, if you have a set-top box for DSS or Direct TV, that is MPEG-2. Also MPEG-2 is the standard for DVD, and so if you have a DVD player, you're watching MPEG-2 video.
DVORAK: Now, is there another... are they going to do an MPEG-3, or...?
DAY: No, there's not really an MPEG-3, although some... if you look at the MPEG-3 spec, MPG-3, the audio part of MPEG has been adopted for audio, you know, downloading audio files over the Internet. There is an MPEG-4, which is a new thing. It's fairly complex, but it has more to do with video streaming, and it's a new standard that will evolve. I think the jury is out as to what products MPEG-4 will be used in at this stage.
DVORAK: When you talk about, you know, the MPEG standard, where does it fit into with high resolution TV? Is there a compatibility issue there? I've always wondered.
DAY: The HDTV is still based around MPEG. There are higher resolutions, and there are higher bit rates. So if you look at DTV or HDTV, they're still based on MPEG...
DVORAK: MPEG-2.
DAY:Yes. But they are at greater resolutions, greater bit rates, but all still based around the same basic video format.
DVORAK: Okay, so we're not dealing with something that's totally incompatible.
DAY:No.
DVORAK: Now, I noticed that there's a never-ending debate about hardware MPEG decoding and software MPEG decoding. What's the difference? I mean, I know what the difference is, but what is the thinking in the industry as to the suitability of each of these technologies, and maybe you should explain the difference.
DAY:Yeah, sure. With today's PC's being up at around 500 megahertz with a Pentium III, they have significant horsepower to do software decoding. So in previous generations of personal computers, when people have a DVD ROM drive, you bought an add-in card, which had hardware MPEG-2. The advantage was it had very high quality, you have the ability to connect up to your television set or your high-fi system, and you also can use the PC for other things at the same time as watching the video because the processor is freed up from doing the software decode. With machines reaching 500 megahertz, now the quality is becoming good doing software decode. The drawbacks of software decode versus hardware are that you can't really use the PC for anything else at the same time because the Pentium is busy doing the decoding. Or if you've got a notebook computer, say, if you're sitting in an airplane and you want to watch a movie, the battery life, because the Pentium uses up a lot of watts from the battery to do the decoding, you may find that you can't watch an entire movie on the plane without recharging the batteries. So the advantages of hardware are really lower power, and the ability to connect up to a TV or a high-fi system, and providing the best quality available.
DVORAK: I noticed when the DVD's first came out there was a lot of, what you would call artifacting. Which is when you're looking at the picture the next thing you know, you see a big square block, or there's something screwy that happens, you know, it's like the image is messed over and it just pops back into place after a frame or two. With the more recent... I have a new Pioneer DVD player, and I don't see this ever anymore. Is there something that's changed?
DAY: Well, you shouldn't ever have seen it. I mean, the spec and the hardware to play back the DVD disks, you know, should not allow you to have video that is breaking or tearing, or breaking up into blocks. So it may have just been that some of the very early DVD players really have not got all the bugs wrung out of them. The DVD spec is actually quite wide, and so you can actually encode the disk at 2 megabits or up to 10 megabits per second. And I think some of the early players struggled with some of the disks and the way that they were encoded.
DVORAK: Now, when you code at a lower bit rate, you get a more compressed file, is that what you end up with?
DAY:That's correct.
DVORAK: So that's why, you know, some of the disks are all... say, you take three disks and they're all totally full, and one of them has a shorter movie on it, but it's still full. Is that because it may be coded at a higher bit rate?
DAY: It could be. It also depends on the content. If you've got a very still image, it doesn't need so many bits to represent it, there's not much movement. If you have an image with a lot of motion, for example, an explosion scene in a movie, that takes a lot of bits to encode it because there are so many changes going on from one frame to another.
DVORAK: Can this bit rate vary during the playing of a movie?
DAY:Absolutely. One of the differences between MPEG-1 and MPEG-2 is that MPEG-2 is a variable bit rate, so it can change. I think the DVD specification goes up to a little bit over 10 megabits per second as a maximum. Most of the time, it would be significantly less than that. But during peak scenes, such as an explosion, the bit rate would increase.
DVORAK:And this is done by hand, when they're actually making the... they're ripping the...?
DAY: No, this...
DVORAK: How's it done?
DAY: It's very complex. It's actually usually done by the...
DVORAK: Because I'm imagining some guy standing, 'Oh, there's an explosion. Let's turn up the bit rate.'
DAY: It's usually done by... the people who are doing the encoding will quite often, for example, be using the C-Cube encoded chip. And that chip through some very complex software and microcode has the ability to look at the video scene and then allocate bits according to what is necessary to make the quality very high. So I'm sure in the studio, they are doing other tricks as well, but there's also a lot of complexity within the encoding chips themselves to create the highest quality image.
DVORAK: So what's next on the agenda for C-Cube?
DAY: Next on the agenda for C-Cube. We've been providing video chips to PC companies, to the broadcast companies, and more recently, to the editing companies. The major trend that is now happening in digital video is the consumer market. In most people's homes at the moment, you have an analog TV, you have a VHS player, and the trend over the next few years are that basically anything you have in your house that is analog will move to some kind of a digital product for higher quality and for convenience reasons. So we are actively working on support for time shifting, which is really to accommodate the next generation of video recording in your house, as well as working towards digital television.
DVORAK: You're talking about devices like the ReplayTV?
DAY: Yeah, that kind of box. Today, with a tape-based VHS player, if you came home, say, and you were 30 minutes into the soccer game you wanted to watch, today, you've got a choice--you either wait for the tape to completely finish the game, which may mean sitting around for another hour, and then you have to rewind and then start watching the game for the beginning; it may be very late before you get to bed. But with a hard disk-based system, using probably MPEG, you can start watching the game from the moment you get home, so it's actually recorded in parallel at the same time that you're watching it. So this is the idea of time shifting, where it really doesn't matter what time the video was broadcast, because you can start watching it at any point, and it will continue recording in parallel.
DVORAK: And you don't think this is just because people watch too much TV that something like this is necessary?
DAY: That might well be the case.
DVORAK:Well, that's kind of depressing. I would take a look, for people out there who want to find out more about C-Cube Microsystems, at www.c-cube.com. We're talking to Chris Day, Senior Director of Marketing for the Kodak Division there in Milpitas, at C-Cube Microsystems. Chris, thanks for being with us today.
DAY: Thank you very much. |
